Open thread on Edwards 5 Radial

[tornitore45]

petertha To be sure, you are referring to an engine with the lifters 15 degree apart, not the Edwards. Althoug I am thinking of offsetting them on my Edward's as well.

 

[tornitore45]

Different subject. I received the pinions, the ring gear ETA was quoted between March and April, still will be here before I need it.

Looking at the 12 teeth pinion, I am not to comfortable boring it to 0.312, there is very metal left to the minor diameter. Granted the 12T will be brazed to the 18 T face to face, but still think is too weak. I plan to bore only 1/3 of the way for a steel bushing to keep center and run on a bronze shaft.

 

[tornitore45]

For one it reduces the poppet-pushrod-rocker angle a possible safety against a pushrod jumping out at Hy-Rev.

On close inspection it does not. The rockers are slanted.

 

[petertha]

If you move the lifters to typical equal offset from center, then you will also have to modify the cam profile in order to preserve Edwards timing phases because the cams are now arriving relative to those new lifter positions. I'm just completing a spreadsheet that calculates these events but here is a sneak peek. The plot with dashed lines represent lifter intersections on my O5. Adding spaced lifters has the effect of increasing valve overlap on the same cam and depending on dimensions can be significant. The Edwards is on center so dashed line would be at single common 360 position, but hopefully you can visualize the net effect.

 

[tornitore45]

Well is not that the cam profile change "shape". Is simply a matter to "rotate" the intake set from the exhaust set as many degrees as the separation of the lifters.
In other words add the same number of degrees to the table of one of the cams.
I suppose that what you meant.
On second thought I have reconsidered offsetting the lifters because in one direction it would increase the angle between rod and rocker and going in the uder direction will cause the an ugly visual crossing of the rods. So I am sticking with plans. I will add raised "Mesas" around the poppet, to increase their guide length and because it may look kool.
On a different note I machined my first part: The rear cover which will act as template for the hole drilling and cutting the crankcase decagon on both parts as an assembly to keep the mating lines flush.
Started the crankcase and run out of room to put the huge volume of stringy swarf.
The ring gear was shipped early on the quoted April date.

 

[petertha]

Yes I should have said cam lobe 'phasing'. There length & ramp stays the same, just altered position along the rim of the cup & relative to indicated TDC. This assumes you make the new displaced lifter axis as extended radii from center of crankshaft. If the lifters were say vertical or positioned different fore aft, that would necessitate modification of lobe shape to achieve the same Edwards timing.

** actually I take that back, the lift would be different to the rockers & valves from current position if you made new push rod locations, so I think you WOULD have to modify the lobes**

I don't get what you mean by crossing of the rods.

 

[tornitore45]

The rods form a visual "V" by separating the poppets in one direction the rods will look more like \ / an open bottom "V" but separating the poppets in the opposite direction (switching them) then the rods will appear like an unsymmetrical "X" crossing visually near the crankcase.

 

[petertha]

Ya off hand I can't think of a good justification for crossing rods like that, at least on a conventional radial. Using the Edwards as an example, the valve angle is influenced by the combustion chamber shape. That then influences the rocker plane angle relative to center plane. A push rod axis more aligned to the rocker plane appears better aligned, at least in that local pivot area. But continue down the push rod to the lifter & cam - the angle between those components becomes worse as a result.

One thing I haven't factored in my spreadsheet is the lifters occur on slightly different fore-aft positions, coincident with the separate intake/exhaust cam planes. So all thing being equal, will also have an effect on timing. Likely its very small so not worth worrying about. Or maybe Edwards has captured some of this in the slightly different timing duration & onset. Its interesting how much 'similar' 4S engines vary in their basic timing just related to breathing.

 

[petertha]

OK I am going to take a crack at this Edwards cam timing. I hope I measured & calculated things correctly. If you see any errors please let me know because radial engine cam timing is not exactly obvious (at least for me). I figure I have to have a clear understanding for my own project & some of you are building the Edwards, so mutually beneficial.

In order to figure out timing, we need to know 5 positions on the cam plate(s): exhaust open (EO), exhaust close (EC), inlet open (IE), inlet closed (IC) and some piston position usually top dead center (TDC) relative to these positions on a common, typically angular coordinate system. There are 2 sets of I/E cam lobes because of 4:1 planetary gear ratio between cam plate & crankshaft. The crankshaft makes 4 rotations per 1 rotation of cam. So in 2 crankshaft rotations a cylinder will see 1 set of I/E lobes. In 4 crankshaft rotations a cylinder will see both (2) sets of I/E cams lobes. Note - other radials like 7-cyl and 9-cyl have different cam layouts, I'm just talking the Edwards here.

We also need to know direction of rotation. The crankshaft turns CCW (counter clockwise) facing front of engine. It drives an external set of intermediate idler gears CW. The idler gear set is joined face to face. One of them drives an internal ring gear, therefore in same direction, CW. The cam plate is attached to the ring gear. So the net result is CCW crankshaft drives a CW cam plate.

There is another bit of complexity which is the position of the cam followers/lifters. If they are aligned to the cylinder TDC center datum line looking at front of engine, then the timing events can be calculated directly from the cam. If the followers are displaced from center, then we need to additionally factor this because the timing is relative to when each I/E follower sees its respective cam bump event. The Edwards happens to be on center.

So to begin, I picked off the 4 cam event angles from the tabular sheet of coordinates from the plans. I think the table is provided mainly to machine the cams lobe profiles. TDC relative to the lobes is not explicitly indicated on the table but it shows up on another drawing. This relationship isn’t super obvious so hopefully I am interpreting correctly.

As it turns out the exhaust & inlet are of different duration and they are phased different relative to one another relative to TDC. This isn’t surprising, many engines have their own unique timing recipe. Unfortunately the Edwards timing isn’t really summarized anywhere in a format that we can readily compare to other engines like ‘exhaust opens at X-deg BBDC & closes at Y-deg ATDC’ and same thing for intake timing. So that was the objective of my spreadsheet, to figure this out & maybe act as a crude design tool. It does not factor valve lift right now, just timing.

 

[petertha]

We only need to concern ourselves with one set of cam lobes because they are symmetric. These angular positions then feed the spreadsheet.

 

[petertha]

Here is what I am calculating. Also a corresponding plot which shows the exhaust/intake crossover timing in simplified duration relative to 360-deg CS TDC position.

 

[tornitore45]

I think I follow your post and agree completely. The drawing do not show specific and detailed instructions on how to phase the cam but rigorous definition are "encoded" in the drawings scattered across several sheets.
I finally noticed the "0 degree TDC" markon the cam ring on PDF page 16.
That say: The screw hole between the lobes must be on the master cylinder axis when the piston is at TDC.
If you notice the cam table tells the Intake Max Lift is at 25 degrees, that means 100 degrees After TDC. In an ideal situation the valve should have full lift mid way on the down stroke. If the piston had true harmonic motion it would be half way down at 90 degrees, so it seem than 100 degrees sound just right since it is standard procedure to delay and extend the intake to account for the air mass inertia.
I have the spread sheet all worked out, and I believe is much the same as yours.
I wish we could exchange Excel files.

 

[Glorfindel]

Great work!!!!

 

[tornitore45]

Petertha if you care to share your e-mail we can exchange spreadsheet and pictures. My e-mail is [email protected]

 

[petertha]

Sounds good. Let me clean up this mess and I will forward. If anyone spots errors, let me know so I don't spread faulty information.
I was kind of struggling to understand timing on my O5 radial for similar reasons, so this work needed to be done.

 

[petertha]

This is not related to the Edwards specifically, but more of an aside. I mentioned that if you separate the cam lifter positions to either side of center (thus widening out the push rods), that would result in different timing, because timing is dictated by when the intake & exhaust lifters 'see' their respective cam bumps. Now in reality we would not just separate the lifter positions & keep the same cam. We would also need to modify the cam lobe angular positions in order to achieve the same open & close timing of the valves themselves relative to TDC. And as mentioned previously, moving the lifters also affects lift geometry, so all this has to be factored in cam design.

So just for interest I applied an arbitrary 7 degrees of separation to the Edwards cam and you can see the resultant shift in timing from the stock scenario. (The sketch is exaggerated just to show). Of course this is a nonsense situation, it just shows the the resultant sensitivity. But I require this offset parameter in the spreadsheet in order to reverse engineer an engine which does have lifter offset in order to compare apples to apples between engines ie. exhaust open close & inlet open close relative to TDC.

 

[petertha]

So after some more head scratching & comparison to other methanol glow 5-cylinder radials with different variations of lifter/pushrod offset, I came to the conclusion that inputting offset into the spreadsheet doesn't serve a valid purpose & potentially confusing. In fact there is a specific methodology to overlay the intake & exhaust lifter positions onto the cam lobe diagram, picking off angles from respective lifter line intersections. The calculations then handles the rest factored relative to TDC and BDC datums. Bottom line is the Edwards values don't change because lifters are on center. I now have more confidence that the tool works. So rather than clutter up this Edwards post talking about other radials, I'll just post the more streamlined spreadsheet calculation & prettied up timing plot. More to come.

 

[tornitore45]

I did the same spreadsheet for the Hoglet engine, a 42 degree V twin cylinders.
It has 3 cams with one cam shared between cylinders. The cams lobes are ad different angles between them and all 4 poppet have different "exit" angles. many variables to keep track.
Basically I made a table associating the position of the crank referred to "Cylinders centerline" and the position of all the cams with their specific angles + a variable offset parameter.
Then I added a column for each cam showing the angular "distance" from the cam lobe to the poppet by adding/subtracting the poppet angular position.
Then added "lift" columns for each poppet when their angular position was within +/- their lift angle. The graph for those was a value of 1 within the lift angle and zero outside.
Then plotted two sinusoid pased +/- 21 degrees representing the piston location.

The entire graph shows square waves identifying the beginning and end of each lift action for all valves and the piston position.

By using different color and different dot mark a very busy diagram was easy to follow. I played with the only variable offset until I got the event where I wanted.

Then I added the position of the two magnets which are locked in phase with the cams to make sure they appeared in the angular range allowed for the sensor to be set and locked.

Is tedious work, but not too difficult, this, of course, has nothing to do with the Edwards but is just to say it is work that should be done for every engine we build in order to fully understand what is going on, and to check the design.
By the way I had to build the cams 5 times, the first time came out perfect but mirror imaged, then I made 3 dumb mistakes messing up 3 blanks, before doing it right.

 

[tornitore45]

Let's focus on the oil passages.
It will be a challenge drilling two holes 0.052 dia 1" long, starting on angled surfaces but plugging the hole with a plug 0.052 dia 0.030 long is impossible

I plan to make a #4-40 aluminum screw with the thread 0.080 long and enough head to slot it.
Tap the hole, plug it and let part of the stud and head hanging out
The screw is just long enough to partially plug the other hole but it also bite crosswise through most of the hole diameter
Drill the radial hole through the installed screw, opening the passage plugged by the screw.

Relieve the Cam Housing to accept the protruding head

 

[Glorfindel]

Let's focus on the oil passages.
It will be a challenge drilling two holes 0.052 dia 1" long, starting on angled surfaces but plugging the hole with a plug 0.052 dia 0.030 long is impossible

I plan to make a #4-40 aluminum screw with the thread 0.080 long and enough head to slot it.
Tap the hole, plug it and let part of the stud and head hanging out
The screw is just long enough to partially plug the other hole but it also bite crosswise through most of the hole diameter
Drill the radial hole through the installed screw, opening the passage plugged by the screw.

Relieve the Cam Housing to accept the protruding head

Until now, i never did small holes. I was a little bit worried, bit i did some holes at 0.069 and it went better than i tought.